AI data center push collides with rising public resistance

Venture firm Eclipse Ventures just pocketed a 17-fold return after its early bets on semiconductor startup Cerebras Systems paid off handsomely with the company's public debut. The firm turned a total investment of 147 million dollars into 2.5 billion at the IPO price of 185 dollars per share. Eclipse partner Lior Susan pointed to this windfall as validation of a long-held view that tying capital to the physical economy, including hardware and manufacturing, would eventually outpace pure software plays.

That thesis sat out of favor for years when Silicon Valley chased enterprise software and SaaS products. Now Susan argues the old software moat has vanished because tools from companies like OpenAI and Anthropic let anyone generate custom code quickly. Public markets appear to agree, with shares of chipmakers like TSMC and Micron reaching record levels as demand for advanced processors climbs.

Yet this surge in AI-related hardware investment collides with deepening public resistance. Recent polling shows broad discontent across age groups and party lines. An Economist and YouGov survey found more than 70 percent of Americans believe artificial intelligence is advancing too fast, including 68 percent of Republicans and 77 percent of Democrats. Negative sentiment toward the technology has climbed from 34 percent three years ago to just over 50 percent today.

Gallup data reveals even starker numbers among younger Americans, with only 18 percent of those ages 14 to 29 expressing hope about AI's direction. A commencement speech this week that praised the technology as the next Industrial Revolution drew loud boos from graduates, underscoring how detached elite optimism often sounds to everyday listeners.

Executives at frontier AI labs and related firms have expressed surprise at these numbers. Some dismiss the concerns outright, insisting the technology's spread is as unstoppable as the internet once was. Others, like the CEO of an AI email assistant company, claim their teams simply do not encounter the negativity reflected in national surveys.

The disconnect matters because the backlash is already translating into business friction. Americans cite fears that AI will eliminate jobs, drive up electricity bills through massive data center demand, concentrate wealth further among a narrow group of owners, and strain the environment. These worries cut across traditional divides and challenge the narrative that rapid deployment faces no real limits.

For investors celebrating gains from chip companies, the physical-world focus may deliver returns precisely because AI systems require enormous amounts of specialized hardware and energy. That same infrastructure, however, amplifies the very costs and disruptions ordinary households now report worrying about most. The result is a widening gap between those positioned to profit from the buildout and those who see their prospects and paychecks placed at risk.